The challenge of environmental history is to explain how and why environments have changed. But it is a demanding task. It is often easier to explain how particular interpretations of the environment formed, why they persisted and how they have been contested and reproduced by administrators, politicians and residents. Accounts of this type cannot always determine whether the interpretation in question was right or wrong. Ultimately, asking how a view came to be, and how it varied or stuck, does not tell us how correct it was. Sometimes no light is shed on whether or not statements made about forests, deserts, lakes, pasture or soil were true or false.
In one sense it is not a useful exercise to discern whether a particular theory portrays the truth about the environment. Groups construct images of the environment and environmental change that work for them which are not really within the realm of Popperian refutation. They exist and persist for a number of reasons in which `hard evidence' or `data' will not necessarily figure (Leach and Mearns, 1996). But discussing perceptions of nature only in terms of their social construction is unsatisfactory--it risks replacing natural reductionism with social reductionism (Entrikin, 1996; Proctor, 1998). Social constructions of nature say things about the environment that can be tested.
We hold that it is possible to test statements about the environment with data if one sets proper caveats to their use and the conclusions drawn. Our difficulties are more practical than theoretical. Sometimes it is not possible to reach firm conclusions from the data available. A major difficulty facing African environmental histories in general is how to cope with a paucity of good data. Speculation is always possible, but rejecting hypotheses may not be.
In this article we confront these dilemmas in the Mkomazi Game Reserve of northern Tanzania. There has been official and local concern over the environmental effects of pastoralism here for over 60 years. The government's concerns, local responses to them and pastoral resistance to central control ensure a wealth of records, claims and opinions about Mkomazi's environment. We have argued that competing claims about Mkomazi's environment should be taken as alternative hypotheses (Brockington and Homewood, 1996). We now wish to examine some of the claims that have been made about Mkomazi's environment. Our goal is to test and if possible refute some of the hypotheses concerning vegetation change, overgrazing, livestock-wildlife interactions and resource management (Table 1).
We first outline the nature of the disputes and the complexity of the different world views brought to bear on Mkomazi. Then we consider the data cited to support ideas about Mkomazi's environment relevant to these hypotheses. Where data have not been cited we assess what evidence exists to support or refute the ideas.
DISPUTES OVER MKOMAZI
Mkomazi's environment has long been contested, and the last fifty years have seen disputes about Mkomazi grow from local to national and ultimately international concerns (Rogers et al., 1999; Brockington, forthcoming). When it was set up in 1951, residence was allowed to some Parakuyo pastoralists who lived in the east of the reserve with a few thousand cattle. But the reserve held excellent pasture and was also used by hunters and honey gatherers, and for wild food and fuel wood. Legal and illegal use of its resources mounted over the next 35 years. By 1984 there were nearly 100,000 cattle living in and around its borders (Homewood et al., 1997). The Department of Wildlife reasserted control over the reserve in 1988, evicting all who lived inside it and prohibiting any resource use. The principal reason given for the eviction of people was that they were destroying the reserve's environment (Mangubuli 1991).
Since it was cleared, Mkomazi has come to international attention both as a conservation triumph and as the focus of human rights concerns. …